Manage Cloud Topics

Starting in v26.1, Redpanda provides Cloud Topics to support multi-modal streaming workloads in the most cost-effective way possible: as a per-topic configuration running mixed latency workloads. While standard Redpanda topics that use local storage or Tiered Storage are ideal for latency-sensitive workloads (for example, for audit logs or analytics), Cloud Topics are optimized for latency-tolerant, high-throughput workloads where cross-AZ networking charges are a major consideration that can become the dominant cost driver at high throughput. These workloads can include observability streams, offline analytics, AI/ML model training data feeds, or development environments that have flexible latency requirements.

Instead of replicating every byte across expensive network links, Cloud Topics leverage durable, inexpensive cloud storage (S3, ADLS, GCS, MinIO) as the primary mechanism to both replicate data and serve it to consumers. This eliminates over 90% of the cost of replicating data over network links in multi-AZ clusters.

The end-to-end latency experienced when using Cloud Topics can range from 500 ms to as high as a few seconds with different object stores. Lower latencies may be achievable in certain environments, but Cloud Topics is optimized for throughput rather than low latency or tightly constrained tail latency. This latency profile is often acceptable for many streaming workloads, and can unlock new streaming use cases that previously were not cost effective.

With Cloud Topics, data from the client is not acknowledged until it is uploaded to object storage. This maintains durability in the face of infrastructure failures, but results in an increase in both produce latency and end-to-end latency, driven by both batching of produced data and the inherent latency of the underlying object store. You should generally expect end-to-end latencies of 1-2 seconds with public cloud stores.